Data Storage; Tiers, not Tears

Glyn Heath, MD of Centiq Ltd, says small companies are struggling as they wrestle with the ball-and-chain of unstructured data

A recent data health-check on a mid-sized organisation found an average of 5,000 different spreadsheets per user.  On another occasion, directors admitted they didn't know what unstructured data was, let alone have a management policy for it.

Businesses are paying over the odds for IT maintenance and management costs because of data storage issues.  Companies are also preventing their key applications and their staff from responding effectively to their customers.  Many companies are not getting a grip on the huge amounts of data that their organisation generates, and so it continues to spiral out of control. 

The problem is growing.  Because of its apparent ease, most users use email to distribute information, generating duplicates and forcing the office manager to impose restrictions on mailbox sizes, in order to cope.  Some businesses have turned to Microsoft SharePoint to help, but in so doing have simply "ported" poorly-managed file systems to a more costly environment and failed to sort the wheat from the chaff.  Small firms accept, generate or simply pass on a flood of Microsoft Office documents, JPEGS, PDFs and video - never mind the emails.  At best, users become inefficient as they struggle to remember where their information resides.  At worst, this user-defined retention policy could be putting businesses at serious risk. 

Left to their own devices, IT departments rarely ask the question, "which data to keep and for how long?" or when they do, get the "all of it forever" response which, without tailored tools, cannot be challenged.  Every industry has regulations, so it can be difficult to organise data retrieval and archiving - so why not carry on as before?

However, this flawed "let sleeping dogs lie" approach causes operational issues:

• Slower information retrieval and questionable quality.  Are critical emails being dumped or saved, but impossible to find?
• Additional IT overheads of extra storage infrastructures.  Small businesses may have to manage one Terabyte (TB) of unstructured data, but it isn't simply a case of buying a £100 drive.  With connectivity, replication, management and back-up and recovery overheads, the associated costs could run into thousands of pounds per year
• 30-40% of a users time could be spent managing (or finding) unstructured data

The person in charge of IT must gain the board's support to understand the critical data that the business holds, where it comes from, and how best to manage it.  They must determine which data is needed immediately (sales enquiry details), which could wait until tomorrow (contract details) and which could wait a few days (pension policy queries).  Some data has to be kept for decades, but when required, it would be needed in seconds.

Nobody finds managing a flow of emails, unwanted files and redundant JPEGs easy.  However this task is made easier by the office manager sitting down with the board and taking a series of steps to assign priorities allowing business data to be tiered.  Once the policy is understood, and rules are defined, the door opens for automation techniques to make these processes more efficient and dependable.

Five key steps could ease the strain:

• Understand what data you hold and where it is
• Present audit findings to the board to clarify current practices and costs to the business.  Make recommendations as to how to improve data value in relation to core business processes
• Establish company-wide data management policies that respect SLAs and meet industry regulations
• Classify, then archive data - according to business value - delete valueless information with confidence
• In the future, plan to define your company's data value against changing core business processes.

For every firm that has had a server or key application fall over, or not come back in time from weekly back-ups because the system was overloaded with data, this data management health check is essential.

Network Computing - Data Storage; Tiers, not Tears